T-Match: Privacy-Preserving Item Matching for Storage-Only RFID Tags

Abstract

Abstract. RFID-based tag matching allows a reader R k to determine whether two tags Ti and Tj store some attributes that jointly fulfill a boolean constraint. The challenge in designing a matching mechanism is tag privacy. While cheap tags are unable to perform any computation, matching has to be achieved without revealing the tags' attributes. In this paper, we present T-MATCH, a protocol for secure and privacy preserving RFID tag matching. T-MATCH involves a pair of tags Ti and Tj, a reader R k , and a backend server S. To ensure tag privacy against R k and S, T-MATCH employs a new technique based on secure two-party computation that prevents R k and S from disclosing tag attributes. For tag privacy against eavesdroppers, each tag Ti in T-MATCH stores an IND-CPA encryption of its attribute. Such an encryption allows R k to update the state of Ti by merely re-encrypting Ti's ciphertext. T-MATCH targets cheap tags that cannot perform any computation, but are only required to store 150 bytes.

High-Speed Dating Privacy-Preserving Attribute Matching for RFID

“…In case an attacker is able to obtain the result of the protocol run, an attack on unlinkability becomes trivial as pointed out in [9]. It is sufficient to have one tag participate in two protocol runs with potentially different tags.…”

“…In 2012, Elkhiyaoui, Blass, and Molva presented a protocol that allows an RFID reader to determine whether two tags store some attributes that jointly fulfill a boolean constraint, without violating the privacy of the tags [9]. They motivate their protocol by considering the transportation of chemicals where safety regulations prohibit the joint transportation of chemicals that might react with each other.…”

“…Additionally, unlike in the T-Match protocol presented in [9], readers do not need to perform any homomorphic encryption operations, or expensive multi-party computations.…”